A quantitative evaluation method for quasi-real-time and refined supervision of soil and water conservation

Abstract

The present invention provides a quantitative evaluation method for quasi-real-time and refined supervision of soil and water conservation, and relates to the field of soil and water conservation technology. The quantitative evaluation method for quasi-real-time and refined supervision of soil and water conservation comprises the following steps: S1. Data acquisition; S2. Data preprocessing; S3. Feature extraction; S4. Model construction; S5. Quantitative evaluation; S6. Result visualization; S7. Supervision decision. By utilizing remote sensing technology and automated data processing procedures, the soil and water conservation conditions in a large area can be quickly acquired and analyzed to achieve quasi-real-time monitoring. For possible risks of soil and water loss, timely warnings can be issued and corresponding measures can be taken to reduce losses. Data acquisition, preprocessing, feature extraction, and model application are all automated, which greatly improves work efficiency. Due to its high efficiency, this method is suitable for soil and water conservation supervision in large areas.

Description

Technical Field

[0001] The present invention relates to the technical field of soil and water conservation, and in particular to a quasi-real-time refined supervision and quantitative evaluation method for soil and water conservation. Background Art

[0002] Soil erosion is a global environmental problem that has serious negative impacts on ecosystems, agricultural production, water resources and human infrastructure. In many areas, soil erosion is becoming increasingly serious due to the combined effects of natural factors (such as terrain undulation, uneven precipitation distribution and soil texture) and human activities (such as agricultural reclamation, urbanization, deforestation and mining). This not only leads to a decline in soil fertility and land degradation, but may also trigger a series of chain reactions such as floods, water pollution and ecological imbalance.

[0003] With global climate change and the intensification of human activities, the problem of soil erosion has become increasingly prominent and has become an important factor affecting the ecological environment and agricultural production. Soil erosion not only leads to land degradation and reduced fertility, but may also cause a series of environmental problems, such as increased sediment content in rivers, siltation in reservoirs, and imbalance in the ecosystem. Therefore, soil and water conservation work has received widespread attention worldwide and has become an important part of the ecological environment and sustainable development.

[0004] However, traditional soil and water conservation supervision methods mainly rely on manual inspections and empirical judgments. This method has problems such as low efficiency, insufficient accuracy and difficulty in covering large areas. At the same time, traditional soil and water conservation supervision methods often rely on regular ground surveys and relatively lagging statistical data. The update cycle of this data acquisition method is long and cannot timely reflect the dynamic changes in soil and water conservation conditions. Therefore, technical personnel in this field provide a quasi-real-time and refined supervision quantitative evaluation method for soil and water conservation to solve the problems raised in the above background technology. Summary of the invention

[0005] 1. Technical issues to be solved

[0006] In view of the shortcomings of the existing technology, the present invention provides a quasi-real-time refined supervision and quantitative evaluation method for soil and water conservation, which solves the problems of low efficiency, insufficient accuracy and difficulty in covering large areas in the existing methods.

[0007] (II) Technical solution

[0008] To achieve the above objectives, the present invention is implemented through the following technical solutions: a quantitative evaluation method for quasi-real-time refined supervision of soil and water conservation, comprising the following steps:

[0009] S1. Data acquisition: use remote sensing satellites and UAV equipment to obtain multi-temporal remote sensing image data of the target area, and collect basic geographic information data such as meteorological data, terrain data, soil data and land use data of the target area;

[0010] S2. Data preprocessing: perform radiation correction, geometric correction and preprocessing of the acquired remote sensing image data to affect fusion, and standardize the basic geographic information data;

[0011] S3. Feature extraction: based on the pre-processed remote sensing image data, extract the characteristic parameters of vegetation index, surface temperature and soil temperature reflecting the soil and water conservation status;

[0012] S4. Model construction: Combine the extracted characteristic parameters and basic geographic information data to build a soil and water conservation status evaluation model, and use machine learning algorithms to train and optimize the model;

[0013] S5. Quantitative evaluation: input the real-time remote sensing image data of the target area into the trained evaluation model to generate quantitative evaluation results of soil and water conservation status;

[0014] S6. Visualize the results. Use GIS technology to spatially express the evaluation results and generate a soil and water conservation status distribution map;

[0015] S7. Supervision decision-making: formulate corresponding soil and water conservation supervision measures and decision-making recommendations based on the evaluation results.

[0016] Preferably, the data acquisition comprises the following steps:

[0017] S1. Acquisition of remote sensing image data: determine the research area and the required remote sensing data source, select appropriate satellite or UAV platform according to the research purpose, access the remote sensing data platform, download remote sensing image data of the target area, determine the time range and frequency of data acquisition according to the climate characteristics and vegetation growth cycle of the research area, download and store remote sensing image data, and ensure the integrity and availability of the data;

[0018] S2. Collection of basic geographic information data: obtain meteorological data of the target area from the meteorological department or relevant database, including temperature, precipitation and wind speed, obtain digital elevation model data for analyzing the terrain characteristics of the area, collect data on soil type, geology and fertility, which can be obtained through soil database or field survey, obtain land use status map, analyze land use type and distribution, including the following calculation formula:

[0019] Normalized Difference Vegetation Index, used to assess vegetation coverage and growth conditions:

[0020]

[0021] Among them, NIR represents the reflectivity of the near-infrared band, and Red represents the reflectivity of the red light band;

[0022] Enhanced Vegetation Index: EVI is an improvement on NDVI, reducing soil background and atmospheric effects;

[0023]

[0024] Among them, Blue represents the reflectivity of the blue light band;

[0025] Surface temperature:

[0026] LST=B(T sensor )+C

[0027] Among them, B(T sensor) is the brightness temperature of the sensor temperature, C is the correction coefficient;

[0028] Soil moisture:

[0029] SM=α×(TB v -TB H )+b

[0030] Among them, TB v and TB H are the brightness temperatures for vertical and horizontal polarizations, respectively, and α and b are empirical coefficients;

[0031] Soil and Water Conservation Index:

[0032] SCI=ω 1 ×NDVI+ω 2 ×LST+ω 3 ×SM+·…

[0033] Among them, ω 1 ,ω 2 ,ω 3 , .... is the weight of each indicator, which can be determined by expert knowledge or machine learning algorithm.

[0034] Preferably, the data preprocessing comprises the following steps:

[0035] S1. Preprocessing of remote sensing image data: first, radiation correction is performed to correct the radiation error caused by sensor characteristics and atmospheric factors, then geometric correction is performed to correct the geometric distortion caused by satellite attitude changes and the rotation of the earth, and finally, impact fusion is performed to fuse the impact data of different spatial resolutions to improve the spatial resolution of the impact. Specifically, the calculation steps include the following:

[0036] Radiometric correction:

[0037]

[0038] Where L is the corrected surface reflectance, L sensor is the radiation brightness received by the sensor, L path is the atmospheric path radiation, T g is the gas permeability, T z is the total atmospheric transmittance, θ s is the solar zenith angle;

[0039] Geometric correction:

[0040] x'=a 0 +a 1 x+a 2 y+a 3 xy+a 4 x 2 +a 5 y 2

[0041] y'=b 0 +b 1 x+b 2 y+b 3 xy+b 4 x 2 +b 5 y 2

[0042] Among them, x and y are the original image coordinates, x' and y' are the corrected impact coordinates, and a 0 , a 1 , ..., a 5 and b 0 , b 1 , ..., b 5 are the polynomial coefficients;

[0043] S2. Preprocess and standardize basic geographic information data, convert the format and projection of data from different sources to ensure data consistency, clean the data, remove redundant data, erroneous data and outliers, and ensure data accuracy and reliability.

[0044] Preferably, the feature extraction comprises the following steps:

[0045] S1. Extract vegetation index: select normalized vegetation index, extract corresponding band data from preprocessed remote sensing image according to the selected vegetation index, apply the calculation formula of vegetation index, calculate the extracted band data, and obtain vegetation index map;

[0046] S2. Extraction of land surface temperature: Select appropriate thermal infrared band data, usually the 6th or 10th band of the Landsat series satellites, apply the land surface temperature inversion algorithm, convert the thermal infrared band data into land surface temperature, and generate a land surface temperature map;

[0047] S3. Soil moisture extraction: Select appropriate microwave remote sensing data, apply soil moisture inversion algorithm, convert microwave remote sensing data into soil moisture, and generate soil moisture map.

[0048] Preferably, the model building comprises the following steps:

[0049] S1. Data preparation: collect and organize various feature data obtained in the feature extraction stage, including vegetation index, surface temperature and soil moisture, as well as basic geographic information data, including topography, soil type and land use, clean and preprocess the data, including removing outliers, filling missing values ​​and data standardization, and divide the data into training sets and test sets, usually using random sampling;

[0050] S2. Model selection: select an appropriate machine learning model based on the research purpose and data characteristics, taking into account the complexity and interpretability of the model, as well as the demand for data volume;

[0051] S3. Model training: Use the training set data to train the selected model, adjust the model parameters to optimize the model performance, and use the cross-validation method to evaluate the model performance to avoid overfitting;

[0052] S4. Model evaluation: Use the test set data to evaluate the trained model, calculate the evaluation indicators, and adjust and optimize the model according to the evaluation results;

[0053] S5. Model application: Apply the trained model to actual data to predict and evaluate the soil and water conservation status, and generate a spatial distribution map of the soil and water conservation status based on the model output results.

[0054] Preferably, the quantitative evaluation comprises the following steps:

[0055] S1. Data preparation: collect and organize the trained models obtained in the model building phase, as well as the real-time remote sensing image data and other relevant data used for evaluation, and pre-process the real-time data, including radiation correction, geometric correction and data standardization, to ensure that the data is consistent with the data format and range used in training the model;

[0056] S2. Model application: input the preprocessed real-time data into the trained model, run the model, and generate the prediction results of soil and water conservation status;

[0057] S3. Result analysis: Analyze the prediction results output by the model, identify the spatiotemporal distribution characteristics of soil and water conservation status, evaluate the current soil and water conservation status based on the analysis results, and determine the areas and issues that need to be focused on;

[0058] S4. Visualize the results. Use GIS software to spatially express the model prediction results, generate soil and water conservation status distribution maps, and make dynamic maps or interactive maps to intuitively display the changing process and trend of soil and water conservation status;

[0059] S5. Supervision decision-making: Based on the quantitative evaluation results, formulate corresponding soil and water conservation supervision measures and decision-making recommendations, and feedback the evaluation results and decision-making recommendations to relevant departments and personnel to guide actual soil and water conservation work.

[0060] Preferably, the result visualization comprises the following steps:

[0061] S1. Data preparation: collect the quantitative evaluation result data output by the model, including the soil and water conservation index value and the specific values ​​of various indicators, prepare geographic information data, such as administrative division maps, topographic maps and land use maps, which will serve as the base map for visualization of the results, and ensure that all data are pre-processed and compatible with the geographic information system software;

[0062] S2. Data import: open the GIS software, import the quantitative evaluation result data and geographic information data into the software, check the integrity and accuracy of the data, and ensure that all data have been correctly imported;

[0063] S3. Spatial data processing: spatially associate the quantitative evaluation result data with the geographic information data to ensure that each evaluation result corresponds to the corresponding geographical location. If necessary, spatial analysis of the data is performed to reveal the spatial distribution characteristics of soil and water conservation conditions;

[0064] S4. Map making: select appropriate map projection and scale to ensure map accuracy and readability; design map symbol system, use different colors, symbols and sizes to represent different levels of soil and water conservation status; add map elements such as legend, scale, compass and title to ensure map information is complete; adjust map layout to optimize map aesthetics and information communication effect.

[0065] (III) Beneficial effects

[0066] The present invention provides a quantitative evaluation method for quasi-real-time and refined supervision of soil and water conservation.

[0067] Beneficial effects:

[0068] 1. In the present invention, by utilizing remote sensing technology and automated data processing procedures, the soil and water conservation conditions in a large area can be quickly acquired and analyzed, and quasi-real-time monitoring can be achieved. For possible risks of soil and water loss, timely warnings can be issued and corresponding measures can be taken to reduce losses.

[0069] 2. In the present invention, data acquisition, preprocessing, feature extraction and model application are all automated, which greatly improves work efficiency. Due to its high efficiency, this method is suitable for soil and water conservation supervision in large-scale areas.

[0070] 3. In the present invention, the high spatial resolution of remote sensing image data enables the evaluation of soil and water conservation status to be refined to a specific geographical location, and combines multiple data sources to achieve a multi-dimensional analysis of soil and water conservation status. DETAILED DESCRIPTION

[0071] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0072] Embodiment 1:

[0073] The embodiment of the present invention provides a method for quantitative evaluation of quasi-real-time and refined supervision of soil and water conservation, comprising the following steps:

[0074] S1. Data acquisition: use remote sensing satellites and UAV equipment to obtain multi-temporal remote sensing image data of the target area, and collect basic geographic information data such as meteorological data, terrain data, soil data and land use data of the target area;

[0075] S2. Data preprocessing: perform radiation correction, geometric correction and preprocessing of the acquired remote sensing image data to affect fusion, and standardize the basic geographic information data;

[0076] S3. Feature extraction: based on the pre-processed remote sensing image data, extract the characteristic parameters of vegetation index, surface temperature and soil temperature reflecting the soil and water conservation status;

[0077] S4. Model construction: Combine the extracted characteristic parameters and basic geographic information data to build a soil and water conservation status evaluation model, and use machine learning algorithms to train and optimize the model;

[0078] S5. Quantitative evaluation: input the real-time remote sensing image data of the target area into the trained evaluation model to generate quantitative evaluation results of soil and water conservation status;

[0079] S6. Visualize the results. Use GIS technology to spatially express the evaluation results and generate a soil and water conservation status distribution map;

[0080] S7. Supervision decision-making: formulate corresponding soil and water conservation supervision measures and decision-making recommendations based on the evaluation results.

[0081] Data acquisition includes the following steps:

[0082] S1. Acquisition of remote sensing image data: determine the research area and the required remote sensing data source, select appropriate satellite or UAV platform according to the research purpose, access the remote sensing data platform, download remote sensing image data of the target area, determine the time range and frequency of data acquisition according to the climate characteristics and vegetation growth cycle of the research area, download and store remote sensing image data, and ensure the integrity and availability of the data;

[0083] S2. Collection of basic geographic information data: obtain meteorological data of the target area from the meteorological department or relevant database, including temperature, precipitation and wind speed, obtain digital elevation model data for analyzing the terrain characteristics of the area, collect data on soil type, geology and fertility, which can be obtained through soil database or field survey, obtain land use status map, analyze land use type and distribution, including the following calculation formula:

[0084] Normalized Difference Vegetation Index, used to assess vegetation coverage and growth conditions:

[0085]

[0086] Among them, NIR represents the reflectivity of the near-infrared band, and Red represents the reflectivity of the red light band;

[0087] Enhanced Vegetation Index: EVI is an improvement on NDVI, reducing soil background and atmospheric effects;

[0088]

[0089] Among them, Blue represents the reflectivity of the blue light band;

[0090] Surface temperature:

[0091] LST=B(T sensor )+C

[0092] Among them, B(T sensor ) is the brightness temperature of the sensor temperature, C is the correction factor;

[0093] Soil moisture:

[0094] SM=α×(TB v -TB H )+b

[0095] Among them, TB v and TB H are the brightness temperatures for vertical and horizontal polarizations, respectively, and α and b are empirical coefficients;

[0096] Soil and Water Conservation Index:

[0097] SCI=ω 1 ×NDVI+ω 2 ×LST+ω 3 ×SM+·…

[0098] Among them, ω 1 ,ω 2 ,ω 3 , .... is the weight of each indicator, which can be determined by expert knowledge or machine learning algorithm.

[0099] Data preprocessing includes the following steps:

[0100] S1. Preprocessing of remote sensing image data: first, radiation correction is performed to correct the radiation error caused by sensor characteristics and atmospheric factors, then geometric correction is performed to correct the geometric distortion caused by satellite attitude changes and the rotation of the earth, and finally, impact fusion is performed to fuse the impact data of different spatial resolutions to improve the spatial resolution of the impact. Specifically, the calculation steps include the following:

[0101] Radiometric correction:

[0102]

[0103] Where L is the corrected surface reflectance, L sensor is the radiation brightness received by the sensor, L path is the atmospheric path radiation, T g is the gas permeability, T z is the total atmospheric transmittance, θ s is the solar zenith angle;

[0104] Geometric correction:

[0105] x'=a 0 +a 1 x+a 2 y+a 3 xy+a 4 x 2 +a 5 y 2

[0106] y'=b 0 +b 1 x+b 2 y+b 3 xy+b 4 x2 +b 5 y 2

[0107] Among them, x and y are the original image coordinates, x' and y' are the corrected impact coordinates, and a 0 , a 1 , ..., a 5 and b 0 , b 1 , ..., b 5 are the polynomial coefficients;

[0108] S2. Preprocess and standardize basic geographic information data, convert the format and projection of data from different sources to ensure data consistency, clean the data, remove redundant data, erroneous data and outliers, and ensure data accuracy and reliability.

[0109] Feature extraction includes the following steps:

[0110] S1. Extract vegetation index: select normalized vegetation index, extract corresponding band data from preprocessed remote sensing image according to the selected vegetation index, apply the calculation formula of vegetation index, calculate the extracted band data, and obtain vegetation index map;

[0111] S2. Extraction of land surface temperature: Select appropriate thermal infrared band data, usually the 6th or 10th band of the Landsat series satellites, apply the land surface temperature inversion algorithm, convert the thermal infrared band data into land surface temperature, and generate a land surface temperature map;

[0112] S3. Soil moisture extraction: Select appropriate microwave remote sensing data, apply soil moisture inversion algorithm, convert microwave remote sensing data into soil moisture, and generate soil moisture map.

[0113] Model building includes the following steps:

[0114] S1. Data preparation: collect and organize various feature data obtained in the feature extraction stage, including vegetation index, surface temperature and soil moisture, as well as basic geographic information data, including topography, soil type and land use, clean and preprocess the data, including removing outliers, filling missing values ​​and data standardization, and divide the data into training sets and test sets, usually using random sampling;

[0115] S2. Model selection: select an appropriate machine learning model based on the research purpose and data characteristics, taking into account the complexity and interpretability of the model, as well as the demand for data volume;

[0116] S3. Model training: Use the training set data to train the selected model, adjust the model parameters to optimize the model performance, and use the cross-validation method to evaluate the model performance to avoid overfitting;

[0117] S4. Model evaluation: Use the test set data to evaluate the trained model, calculate the evaluation indicators, and adjust and optimize the model according to the evaluation results;

[0118] S5. Model application: Apply the trained model to actual data to predict and evaluate the soil and water conservation status, and generate a spatial distribution map of the soil and water conservation status based on the model output results.

[0119] The quantitative evaluation includes the following steps:

[0120] S1. Data preparation: collect and organize the trained models obtained in the model building phase, as well as the real-time remote sensing image data and other relevant data used for evaluation, and pre-process the real-time data, including radiation correction, geometric correction and data standardization, to ensure that the data is consistent with the data format and range used in training the model;

[0121] S2. Model application: input the preprocessed real-time data into the trained model, run the model, and generate the prediction results of soil and water conservation status;

[0122] S3. Result analysis: Analyze the prediction results output by the model, identify the spatiotemporal distribution characteristics of soil and water conservation status, evaluate the current soil and water conservation status based on the analysis results, and determine the areas and issues that need to be focused on;

[0123] S4. Visualize the results. Use GIS software to spatially express the model prediction results, generate soil and water conservation status distribution maps, and make dynamic maps or interactive maps to intuitively display the changing process and trend of soil and water conservation status;

[0124] S5. Supervision decision-making: Based on the quantitative evaluation results, formulate corresponding soil and water conservation supervision measures and decision-making recommendations, and feedback the evaluation results and decision-making recommendations to relevant departments and personnel to guide actual soil and water conservation work.

[0125] Result visualization includes the following steps:

[0126] S1. Data preparation: collect the quantitative evaluation result data output by the model, including the soil and water conservation index value and the specific values ​​of various indicators, prepare geographic information data, such as administrative division maps, topographic maps and land use maps, which will serve as the base map for visualization of the results, and ensure that all data are pre-processed and compatible with the geographic information system software;

[0127] S2. Data import: open the GIS software, import the quantitative evaluation result data and geographic information data into the software, check the integrity and accuracy of the data, and ensure that all data have been correctly imported;

[0128] S3. Spatial data processing: spatially associate the quantitative evaluation result data with the geographic information data to ensure that each evaluation result corresponds to the corresponding geographical location. If necessary, spatial analysis of the data is performed to reveal the spatial distribution characteristics of soil and water conservation conditions;

[0129] S4. Map making: select appropriate map projection and scale to ensure map accuracy and readability; design map symbol system, use different colors, symbols and sizes to represent different levels of soil and water conservation status; add map elements such as legend, scale, compass and title to ensure map information is complete; adjust map layout to optimize map aesthetics and information communication effect.

[0130] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A quantitative evaluation method for quasi-real-time and refined supervision of soil and water conservation, characterized by: The following steps are involved: S1. Data acquisition: use remote sensing satellites and UAV equipment to obtain multi-temporal remote sensing image data of the target area, and collect basic geographic information data such as meteorological data, terrain data, soil data and land use data of the target area; S2. Data preprocessing: perform radiation correction, geometric correction and preprocessing of the acquired remote sensing image data to affect fusion, and standardize the basic geographic information data; S3. Feature extraction: based on the pre-processed remote sensing image data, extract the characteristic parameters of vegetation index, surface temperature and soil temperature reflecting the soil and water conservation status; S4. Model construction: Combine the extracted characteristic parameters and basic geographic information data to build a soil and water conservation status evaluation model, and use machine learning algorithms to train and optimize the model; S5. Quantitative evaluation: input the real-time remote sensing image data of the target area into the trained evaluation model to generate quantitative evaluation results of soil and water conservation status; S6. Visualize the results. Use GIS technology to spatially express the evaluation results and generate a soil and water conservation status distribution map; S7. Supervision decision-making: formulate corresponding soil and water conservation supervision measures and decision-making recommendations based on the evaluation results.

2. According to claim 1, a quasi-real-time refined supervision and quantitative evaluation method for soil and water conservation is characterized by: The data acquisition comprises the following steps: S1. Acquisition of remote sensing image data: determine the research area and the required remote sensing data source, select appropriate satellite or UAV platform according to the research purpose, access the remote sensing data platform, download remote sensing image data of the target area, determine the time range and frequency of data acquisition according to the climate characteristics and vegetation growth cycle of the research area, download and store remote sensing image data, and ensure the integrity and availability of the data; S2. Collection of basic geographic information data: obtain meteorological data of the target area from the meteorological department or relevant database, including temperature, precipitation and wind speed, obtain digital elevation model data for analyzing the terrain characteristics of the area, collect data on soil type, geology and fertility, which can be obtained through soil database or field survey, obtain land use status map, analyze land use type and distribution, including the following calculation formula: Normalized Difference Vegetation Index, used to assess vegetation coverage and growth conditions: Among them, NIR represents the reflectivity of the near-infrared band, and Red represents the reflectivity of the red light band; Enhanced Vegetation Index: EVI is an improvement on NDVI, reducing soil background and atmospheric effects; Among them, Blue represents the reflectivity of the blue light band; Surface temperature: LST=B(T sensor )+C Among them, B(T sensor ) is the brightness temperature of the sensor temperature, C is the correction factor; Soil moisture: SM=α×(TB v -TB H )+b Among them, TB v and TB H are the brightness temperatures for vertical and horizontal polarizations, respectively, and α and b are empirical coefficients; Soil and Water Conservation Index: SCI=ω1×NDVI+ω2×LST+ω3×SM+·… Among them, ω1, ω2, ω3, .... are the weights of each indicator, which can be determined by expert knowledge or machine learning algorithms.

3. The method for quasi-real-time and refined supervision and quantitative evaluation of soil and water conservation according to claim 1 is characterized by: The data preprocessing comprises the following steps: S1. Preprocessing of remote sensing image data: first, radiation correction is performed to correct the radiation error caused by sensor characteristics and atmospheric factors, then geometric correction is performed to correct the geometric distortion caused by satellite attitude changes and the rotation of the earth, and finally, impact fusion is performed to fuse the impact data of different spatial resolutions to improve the spatial resolution of the impact. Specifically, the calculation steps include the following: Radiometric correction: Where L is the corrected surface reflectance, L sensor is the radiation brightness received by the sensor, L path is the atmospheric path radiation, T g is the gas permeability, T z is the total atmospheric transmittance, θ s is the solar zenith angle; Geometric correction: <h2 style=";text-align:left;direction:ltr">x' = a0 + a1x + a2y + a3xy + a4x<h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> +a5y<h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr">y' = b0 + b1x + b2y + b3xy + b4x<h2 style=";text-align:left;direction:ltr"> 2 <h2 style=";text-align:left;direction:ltr"> +b5y<h2 style=";text-align:left;direction:ltr"> 2 Where x and y are the original image coordinates, x' and y' are the corrected impact coordinates, a0, a1, ..., a5 and b0, b1, ..., b5 are polynomial coefficients; S2. Preprocess and standardize basic geographic information data, convert the format and projection of data from different sources to ensure data consistency, clean the data, remove redundant data, erroneous data and outliers, and ensure data accuracy and reliability.

4. The method for quasi-real-time and refined supervision and quantitative evaluation of soil and water conservation according to claim 1 is characterized by: The feature extraction comprises the following steps: S1. Extract vegetation index: select normalized vegetation index, extract corresponding band data from preprocessed remote sensing image according to the selected vegetation index, apply the calculation formula of vegetation index, calculate the extracted band data, and obtain vegetation index map; S2. Extraction of land surface temperature: Select appropriate thermal infrared band data, usually the 6th or 10th band of the Landsat series satellites, apply the land surface temperature inversion algorithm, convert the thermal infrared band data into land surface temperature, and generate a land surface temperature map; S3. Soil moisture extraction: Select appropriate microwave remote sensing data, apply soil moisture inversion algorithm, convert microwave remote sensing data into soil moisture, and generate soil moisture map.

5. The method for quasi-real-time and refined supervision and quantitative evaluation of soil and water conservation according to claim 1 is characterized by: The model construction includes the following steps: S1. Data preparation: collect and organize various feature data obtained in the feature extraction stage, including vegetation index, surface temperature and soil moisture, as well as basic geographic information data, including topography, soil type and land use, clean and preprocess the data, including removing outliers, filling missing values ​​and data standardization, and divide the data into training sets and test sets, usually using random sampling; S2. Model selection: select an appropriate machine learning model based on the research purpose and data characteristics, taking into account the complexity and interpretability of the model, as well as the demand for data volume; S3. Model training: Use the training set data to train the selected model, adjust the model parameters to optimize the model performance, and use the cross-validation method to evaluate the model performance to avoid overfitting; S4. Model evaluation: Use the test set data to evaluate the trained model, calculate the evaluation indicators, and adjust and optimize the model according to the evaluation results; S5. Model application: Apply the trained model to actual data to predict and evaluate the soil and water conservation status, and generate a spatial distribution map of the soil and water conservation status based on the model output results.

6. The method for quasi-real-time and refined supervision and quantitative evaluation of soil and water conservation according to claim 1 is characterized by: The quantitative evaluation includes the following steps: S1. Data preparation: collect and organize the trained models obtained in the model building phase, as well as the real-time remote sensing image data and other relevant data used for evaluation, and pre-process the real-time data, including radiation correction, geometric correction and data standardization, to ensure that the data is consistent with the data format and range used in training the model; S2. Model application: input the preprocessed real-time data into the trained model, run the model, and generate the prediction results of soil and water conservation status; S3. Result analysis: Analyze the prediction results output by the model, identify the spatiotemporal distribution characteristics of soil and water conservation status, evaluate the current soil and water conservation status based on the analysis results, and determine the areas and issues that need to be focused on; S4. Visualize the results. Use GIS software to spatially express the model prediction results, generate soil and water conservation status distribution maps, and make dynamic maps or interactive maps to intuitively display the changing process and trend of soil and water conservation status; S5. Supervision decision-making: Based on the quantitative evaluation results, formulate corresponding soil and water conservation supervision measures and decision-making recommendations, and feedback the evaluation results and decision-making recommendations to relevant departments and personnel to guide actual soil and water conservation work.

7. The method for quasi-real-time and refined supervision and quantitative evaluation of soil and water conservation according to claim 1 is characterized by: The result visualization comprises the following steps: S1. Data preparation: collect the quantitative evaluation result data output by the model, including the soil and water conservation index value and the specific values ​​of various indicators, prepare geographic information data, such as administrative division maps, topographic maps and land use maps, which will serve as the base map for visualization of the results, and ensure that all data are pre-processed and compatible with the geographic information system software; S2. Data import: open the GIS software, import the quantitative evaluation result data and geographic information data into the software, check the integrity and accuracy of the data, and ensure that all data have been correctly imported; S3. Spatial data processing: spatially associate the quantitative evaluation result data with the geographic information data to ensure that each evaluation result corresponds to the corresponding geographic location. If necessary, spatial analysis of the data is performed to reveal the spatial distribution characteristics of soil and water conservation conditions; S4. Map making: select appropriate map projection and scale to ensure map accuracy and readability; design map symbol system, use different colors, symbols and sizes to represent different levels of soil and water conservation status; add map elements such as legend, scale, compass and title to ensure map information is complete; adjust map layout to optimize map aesthetics and information communication effect.

Images